Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Polygenic associations with clinical and neuropathological trait heterogeneity across TDP-43 proteinopathies

  • Barbara E. Spencer,

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • David J. Irwin,

    Roles Funding acquisition, Investigation, Project administration, Supervision, Writing – review & editing

    Affiliation Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • Vivianna M. Van Deerlin,

    Roles Writing – review & editing

    Affiliation Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • EunRan Suh,

    Roles Writing – review & editing

    Affiliation Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • Edward B. Lee,

    Roles Methodology, Writing – review & editing

    Affiliation Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • Lauren Elman,

    Roles Investigation, Writing – review & editing

    Affiliation Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • Colin C. Quinn,

    Roles Investigation, Writing – review & editing

    Affiliation Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • Defne Amado,

    Roles Investigation, Writing – review & editing

    Affiliation Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • Michael Baer,

    Roles Investigation, Writing – review & editing

    Affiliation Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • Murray Grossman ,

    Roles Investigation, Writing – review & editing

    † Deceased

    Affiliation Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • David A. Wolk,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯
  • Corey T. McMillan

    Roles Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Visualization, Writing – original draft, Writing – review & editing

    cmcmilla@pennmedicine.upenn.edu

    Affiliation Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America

    ⨯

Abstract

TDP-43 proteinopathies, including amyotrophic lateral sclerosis (ALS), frontotemporal lobar degeneration with TDP-43 (FTLD-TDP), and limbic-predominant age-related TDP-43 encephalopathy, encompass a spectrum of clinical and neuropathological traits. Despite mounting evidence for shared genetic risk across TDP-43 proteinopathies, the modifiers of individual-level traits are unknown. We aimed to identify polygenic contributions to trait heterogeneity across TDP-43 proteinopathies. We used weighted correlation analysis of GWAS summary statistics for ALS, FTLD-TDP, and hippocampal sclerosis of aging (HS-Aging) to identify data-driven clusters of highly correlated single nucleotide polymorphisms (SNPs). We performed gene ontology enrichment analysis for each identified cluster. We derived cluster-specific polygenic scores and evaluated their association with clinical and neuropathological traits in an independently evaluated sample of individuals who met neuropathological and/or genetic criteria for FTLD-TDP or ALS (n = 260). We identified 5 distinct data-driven clusters, including 3 GWAS phenotype-specific clusters (FTLD-TDP, ALS, HS-Aging) and 2 clusters representing the overlap between a pair of GWAS phenotypes (ALS-FTLD and FTLD-HS). Pathway analysis revealed biologically meaningful associations including distinct GWAS phenotype-specific processes within clusters. Cluster-specific ALS and FTLD-TDP polygenic risk each associated with individual-level clinical traits, even within the context of autosomal dominant mutation carriers, where higher ALS polygenic risk associated with neuromuscular impairment and higher FTLD-TDP polygenic risk associated with cognitive-behavioral impairment. Moreover, higher FTLD-TDP polygenic risk associated with higher TDP-43 burden within characteristic FTLD-TDP brain regions. We suggest that there are polygenic modifiers of clinical and neuropathological traits across TDP-43 proteinopathies that may contribute to individual-level differences, including likelihood for developing FTLD or ALS.

Citation: Spencer BE, Irwin DJ, Van Deerlin VM, Suh E, Lee EB, Elman L, et al. (2025) Polygenic associations with clinical and neuropathological trait heterogeneity across TDP-43 proteinopathies. PLoS One 20(12): e0338398. https://doi.org/10.1371/journal.pone.0338398

Editor: Vijay Kumar, Amity University, INDIA

Received: July 16, 2025; Accepted: November 21, 2025; Published: December 30, 2025

Copyright: © 2025 Spencer et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All summary statistic data was accessed from prior GWAS. Genotype, diagnosis, and neuropathological data used in this manuscript is restricted due to sensitive patient information including genetic risk for neurodegenerative disease. Therefore, access is only granted upon reasonable request to the Penn Neurodegeneration Data Sharing Committee (PNDSC) that will be reviewed by a panel of key stakeholders and only granted following completion of an institutionally certified Data Use Agreement (DUA). To make a data request please use the following webform: https://www.pennbindlab.com/data-sharing.

Funding: National Institutes of Health grant F32AG079618 (BES), P01AG066597 and RF1NS145263 (CTM), NS109260 (DJI), P30AG072979 (DAW), Penn Institute on Aging (CTM, DJI, DAW), and The DeCrane Family Fund for Primary Progressive Aphasia (CTM, DJI). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: All authors declare they have no relevant competing interests.

Introduction

TAR DNA-binding protein ~ 43kDa (TDP-43) inclusions are the pathological hallmark of frontotemporal lobar degeneration with TDP-43 (FTLD-TDP), > 98% of amyotrophic lateral sclerosis (ALS), and ~90% of hippocampal sclerosis of aging (HS-Aging) [1–4]. Despite shared neuropathological traits, individuals with TDP-43 proteinopathies present with heterogeneous clinical traits including neuromuscular impairments (i.e., ALS) and cognitive-behavioral impairments (i.e., FTLD-TDP). TDP-43 pathology and neurodegeneration are observed in characteristic neuroanatomical regions that correlate with clinical traits. Some individuals present with both neuromuscular and cognitive-behavioral impairment (i.e., ALS-FTD) which, together with the shared neuropathological traits support the notion that FTLD-TDP and ALS are part of a clinicopathologic spectrum [5]. Alternatively, individuals with HS-Aging typically present with an amnestic syndrome with an older age of onset [6–8]. Many efforts have focused on the identification of genetic variants that predispose individuals to FTLD-TDP, ALS, or HS-Aging. However, the contribution of genotypes to the shared and disparate clinical and neuropathologic traits observed across TDP-43 proteinopathies is unclear.

Large scale case-control genome wide association studies (GWAS) have focused on the identification of single nucleotide polymorphisms (SNPs) associated with risk for ALS [9,10], FTLD-TDP [11], or HS-Aging [3,12]. Several large-scale GWAS established ALS as a complex, polygenic disease [9,10,13]. The first GWAS in FTLD-TDP cases identified three SNPs spanning TMEM106B that reached genome-wide significance and several studies have validated the importance of these loci in FTLD pathogenesis [14–19]. GWAS of HS-Aging have also implicated several SNPs, including those in TMEM106B [3,12]. Additional efforts have aimed to uncover a shared genetic basis to explain the biological mechanisms that drive susceptibility in FTLD-TDP and ALS. A meta-analysis that jointly analyzed ALS and FTLD-TDP cases identified significant associations in UNC13A [20] which had previously only been identified as a susceptibility SNP for ALS and has since been implicated in an FTLD-TDP GWAS [21]. Another study used a genetic enrichment approach that found an enrichment of ALS as a function of increased association with FTLD-TDP, indicating these phenotypes are genetically related [22]. However, conjunction analyses performed to identify which SNPs were jointly associated with increased risk for ALS and FTLD-TDP only revealed a SNP that is used as a surrogate marker for C9orf72 repeat expansions, which can cause FLTD-TDP and/or ALS even within the same family [23,24].

Despite mounting evidence for shared genetic risk, considerably less effort has focused on the modifiers of individual-level traits. Rather than focusing only on overlapping risk factors, identifying disparate risk alleles may reveal specific neuroanatomic and/or cellular vulnerabilities that contribute to our understanding of trait-specific drivers. Using existing GWAS summary statistics with novel bioinformatics strategies, we aimed to identify polygenic contributions to individual-level clinical and neuropathological traits.

We compared the relative associations of case-control GWAS summary statistics for ALS, FTLD-TDP, and HS-aging to identify data-driven clusters of highly correlated SNPs that may contribute to clinical and neuropathological trait heterogeneity across these neurodegenerative diseases. We identified the biological pathways associated with these clusters. We then related these cluster SNPs to individual-level clinical and neuropathological traits within a neuropathologically and/or genetically defined sample of ALS and FTLD-TDP cases, including the presence of neuromuscular impairment and/or cognitive-behavioral impairment and the burden of TDP-43 pathology across brain regions, in a well-characterized sample of more than 250 individuals with ALS and/or FTLD-TDP, through the construction of cluster-specific polygenic risk scores (PRS). We identified clusters of genotypic contributions to selective vulnerability in the anatomic distribution of TDP-43 associated disease, and, for the first time, established potential genetic contributions to risk for either neuromuscular or cognitive-behavioral impairments within individuals with C9orf72 repeat expansions.

Materials and methods

Cluster identification and annotation

Summary statistic data.

To derive clusters of associated SNPs, we performed weighted correlation analysis on publicly-available summary statistics for reported case-control ALS [9], FTLD-TDP [11], and HS-Aging [12] (which is enriched for TDP-43 [2–4]) GWAS. The ALS GWAS included a combined discovery sample of 20,806 sporadic and familial ALS cases and 59,804 controls. 12,577 cases and 23,475 controls were included from summary statistics downloaded from van Rheenen et al [13]. The remaining 8,229 cases and 36,329 controls were of non-Hispanic white race/ethnicity and were diagnosed with ALS according to the El Escorial criteria [25]. The FTLD-TDP GWAS included 515 FTLD-TDP individuals of European descent with dementia (with or without MND) and either a pathological diagnosis of FTLD-TDP (n = 499) or, in the case of living individuals, a pathogenic GRN mutation (n = 16) and 2,509 controls. The HS-Aging GWAS evaluated 310 cases of HS-Aging as defined by presence vs. absence at pathological evaluation and 2,576 controls.

To harmonize summary statistics across GWAS, we only evaluated SNPs that were commonly genotyped across GWAS. FTLD-TDP summary statistics were transformed from NCBI36/hg18 into GRCh37/hg19 using liftOver [26] for comparison with the ALS and HS-Aging summary statistics. Following alignment, we identified 494,417 SNPs commonly genotyped across these GWAS summary statistics. GWAS statistics for each of the three studies were transformed into z-scores to reduce the effects of sample size differences, and only the absolute value of the z-score was used for analysis. For the ALS GWAS, z-scores were calculated as the effect divided by the standard error. For the FTLD-TDP and HS-Aging GWAS, z-scores were calculated as |z| = |Φ − 1(p/2)| where Φ − 1 is the inverse cumulative distribution function of the normal distribution, and p is the p-value. The top 1% of SNPs with the strongest associations to any GWAS phenotype (by |z|, n = 4,945) were selected for analysis (S1 Fig).

Weighted correlation analysis.

The WGCNA package in R [27] was used to perform weighted correlation analysis. This approach uses the patterns of associations of SNPs with the respective GWAS phenotypes to identify correlations between SNPs and construct clusters based on those correlations. First, an adjacency matrix is constructed using the strength of the association between each SNP and each GWAS phenotype. From the adjacency matrix values, a topological overlap matrix is created to alleviate the effect of noise, and this is converted into a distance matrix which is used as the basis for clustering. Clusters are identified with unsupervised clustering (i.e., without the use of a priori defined gene or SNP sets), specifically hierarchical clustering, along with dynamic tree cut.

As only SNPs that were common across GWAS were considered for analysis, there were 0 SNPs removed due to missingness. Clusters were constructed with a minimum size of 30 SNPs. Clusters whose distance was less than 0.25 were merged. To test for robustness, we created 100 samples, each containing a random 80% of the SNPs selected for weighted correlation analysis, without replacement (S2 Fig). For each resample, weighted correlation analysis was performed as in the full sample.

After clusters were identified, multialleleic SNPs and SNPs with > 10% sample missingness in the individual level genotype data (see Genotype Data) were removed, and SNPs were pruned within clusters using an R2 threshold of 0.3 with LDlinkR [28].

Cluster annotation.

To annotate, cluster SNPs were mapped to overlapping genes using the biomaRt package in R [29], specifically ENSEMBL_MART_SNP. A gene ontology enrichment analysis for biological processes was performed separately for each cluster using gprofiler2 [30].

Enrichment was determined by overrepresentation test, which calculates an expected value by matching the proportion of genes in a particular annotation data category in the human genome reference list to the number of genes provided and calculates a fold enrichment value by comparing the true number of genes in a particular annotation data category to the expected value. A p-value was calculated for each term with the hypergeometric test followed by correction for multiple testing using the False Discovery Rate (FDR), calculated by the Benjamini-Hochberg procedure. Pathways that met FDR p < .05 were examined for each cluster.

Individual- level data in a neuropathologically and/or genetically defined sample

Case selection.

To ensure a TDP-43 proteinopathy, we selected 260 individuals who met neuropathological criteria for FTLD-TDP and/or ALS (n = 174) or had a pathogenic mutation associated with TDP-43 (n = 135, including 49 who also met neuropathological criteria) to include for analysis, described in detail below. SOD1 ALS cases and VCP tauopathy cases were excluded to focus on TDP-43 proteinopathy. Informed written consent was obtained for all participants through a procedure approved by an Institutional Review Board convened at the University of Pennsylvania.

Genotype data.

Individual level imputed genotype data, derived from a variety of prior GWAS array genotyping chips, was available in the Penn Integrated Neurodegenerative Disease Database [31]. To harmonize data, we performed imputation of allele dosages on samples passing QC using the Michigan Imputation Server [32] with the TOPMed reference panel based on>97K sequenced samples [33]. Prior to genotype analyses or imputation, all genotype data were QCed using standard metrics including but not limited to sex discrepancy, sample relatedness, heterozygosity, and Hardy-Weinberg equilibriu. [34].

Genetic sequencing.

DNA was extracted from peripheral blood or frozen brain tissue following the manufacturer’s protocols (QuickGene DNA whole blood kit (Autogen) for blood, and QIAamp DNA Mini Kit (Qiagen) for brain tissue). All individuals were tested for C9orf72 hexanucleotide repeat expansions using a modified repeat-primed PCR as previously described [35] and screened for mutations associated with FTD and/or ALS using whole exome/genome sequencing, and/or a custom targeted multi neurodegenerative disease sequencing panel [36] which included SOD1 [37], TBK1 [38], TARDBP [39], VCP [40], and GRN [41,42]). The sequencing data was analyzed using Geneticist Assistant software (Soft Genetics, State College, PA).

Polygenic risk score calculation.

The weighted correlation analysis identified 5 clusters of correlated SNPs (see Data-driven clusters of correlated SNPs identified using summary statistics). We limited our analysis to clusters associated with a single GWAS phenotype (C1 ALS, C2 FTLD-TDP, and C3 HS-Aging) in an effort to focus on GWAS phenotype-specific drivers. Due to a lack of directionality information in the provided summary statistics, we were unable to calculate a PRS for C3 HS-Aging.

Cluster-specific PRS were constructed for C1 ALS and C2 FTLD-TDP as the weighted sum of alleles per SNP per cluster for each individual, using z-score transformed associations with the respective GWAS phenotype, ALS or FTLD-TDP. The effect allele was labelled as ‘Allele1’ in the ALS summary statistics and was the minor allele in the FTLD-TDP summary statistics. Here, we assigned minor allele using the European 1,000 Genomes sample allele frequencies. The direction of the weight for each SNP was determined by whether the dose of the effect (direction consistent with summary statistic) or non-effect (direction flipped) allele was provided in the individual-level data.

Each PRS was centered using the European 1,000 Genomes sample allele frequencies and mean scaled. Missing allele dosages were replaced with the European 1,000 Genomes sample average. SNPs without allele frequency information were removed. The C1 ALS PRS included 1021 SNPs and the C2 FTLD-TDP PRS included 1013 SNPs. Individuals with >10% SNP missingness for either PRS were removed. This resulted in the removal of 130 individuals.

PRS were tested for association with clinical and neuropathological traits using one-way ANOVAs controlling for age at last contact and sex. Some individuals in the sample were included in a prior ALS or FTLD-TDP GWAS for which summary statistics were used to generate the clusters (Table 1). Therefore, inclusion in a prior ALS or FTLD-TDP GWAS was additionally controlled for in all ANOVAs.

thumbnail
Download:
Table 1. Sample Characteristics.

https://doi.org/10.1371/journal.pone.0338398.t001

Clinical traits.

Within known TDP-43 proteinopathy cases, individuals were characterized based on clinical traits of neuromuscular impairment (MND-only), cognitive-behavioral impairment (Cog-only), and any evidence of a combination of impairment (Cog + MND) at any point in the natural history of disease determined through clinical and research chart review.

Neuropathological traits.

Detailed neuropathological assessments were performed using established and uniform methods of fixation, tissue processing, IHC with well-characterized antibodies, and current neuropathological criteria, as has been described in detail elsewhere [36,43,44]. Briefly, 15 brain regions from one hemisphere, alternating right and left at random, are routinely sampled at autopsy, formalin-fixed, and processed for immunohistochemical staining using 1D3 for phosphorylated TDP-43. Each brain region was semi-quantitatively scored for the burden of TDP-43 (0, absent; 0.5, rare; 1, mild; 2, moderate; 3, severe). ALS was diagnosed based on the presence of upper and lower motor neuron degeneration, and TDP-43 subtypes were determined based on TDP-43 immunohistochemistry as previously described [45].

Results

Data-driven clusters of correlated SNPs identified using summary statistics

Weighted correlation network analysis identified a total of 5 clusters of correlated SNPs associated with either a single GWAS phenotype (accordingly referred to as C1 ALS, C2 FTLD-TDP, and C3 HS-Aging from now on) or a pair of GWAS phenotypes (accordingly referred to as C4 ALS-FTLD and C5 FTLD-HS from now on). Of possible combinations, we notably did not observe a cluster associated with ALS-HS. Cluster assignment was robust, particularly for SNPs associated with a single GWAS phenotype where >97% of SNPs were assigned to the same cluster (C1 ALS, C2 FTLD-TDP, or C3 HS-Aging) when weighted correlation analysis was performed 100 times, each time including a random 80% of the 4,945 SNPs (S2 Fig). After clusters were identified, SNPs were LD pruned within clusters to reduce redundancy (S3 Fig). After pruning, C1 ALS contained 1,025 SNPs, C2 FTLD-TDP contained 1,021 SNPs, C3 HS-Aging contained 646 SNPs, C4 ALS-FTLD contained 80 SNPs, and FTLD-HS contained 72 SNPs (Fig 1, S1 File). In the remaining sections we identify the biological pathways associated with these cluster SNPs and test for individual-level associations between cluster-specific polygenic risk and clinical and neuropathological traits of individuals who met neuropathological and/or genetic criteria for FTLD-TDP or ALS.

thumbnail
Download:
Fig 1. Weighted correlation analysis revealed clusters of correlated SNPs across GWAS phenotypes associated with TDP-43 proteinopathies.

(A) Each cluster SNP is plotted by relative association (by |z|) to FTLD-TDP and ALS. For each cluster, (B) C1 ALS, (C) C2 FTLD-TDP, (D) C3 HS-Aging, (E) C4 ALS-FTLD, and (F) C5 FTLD-HS, radar plots show the relative association of each SNP within that cluster across FTLD-TDP, ALS, and HS-Aging GWAS. Across panels, each dot represents a SNP in C1 ALS (red), C2 FTLD-TDP (blue), C3 HS-Aging (yellow), C4 ALS-FTLD (purple), or C5 FTLD-HS (green). Dashed lines connect a given SNP in each radar plot.

https://doi.org/10.1371/journal.pone.0338398.g001

Pathway analysis revealed biologically plausible associations for cluster SNPs

Beyond individual SNPs, it is likely that biological pathways are involved in the development of complex neurodegenerative diseases. To investigate the biological pathway associations of each of our identified clusters, we performed a gene ontology enrichment analysis for each cluster using gprofiler2 [30]. Enrichment was determined by overrepresentation test, which calculates an expected value by matching the proportion of genes in a particular annotation data category in the human genome reference list to the number of genes provided. Cluster-specific overrepresented pathways suggest the identified clusters are biologically meaningful (FDR p < .05, Fig 2 and S2 File). Notably, C1 ALS uniquely included pathways related to motor neurons and spinal cord, while neurogenesis-related terms were among the top overrepresented pathways in C3 HS-Aging. Additionally, many pathways that were common across clusters related to nervous system processes that may be broadly implicated across neurodegenerative diseases (S4 Fig). Together, these pathway associations support the presence of differentiated and biologically meaningful genetic contributions within the identified clusters. Further, they reinforce the potential for shared susceptibility across diseases, while pointing to potential disparate drivers of ALS, FTLD-TDP, and HS-Aging.

thumbnail
Download:
Fig 2. Derived clusters have biologically plausible pathway associations.

The top 20 overrepresented pathways by fold enrichment that met FDR-adjusted p < .05 are displayed for (A) C1 ALS, (B) C2 FTLD-TDP, and (C) C3 HS-Aging.

https://doi.org/10.1371/journal.pone.0338398.g002

Cluster-specific polygenic risk associated with individual-level clinical and neuropathological trait heterogeneity

While the identified data-driven clusters suggest GWAS phenotype-specific SNPs which may contribute to our understanding of clinical and neuropathological trait heterogeneity across TDP-43 proteinopathies, these summary-level results alone cannot address individual-level risk. We focused on these GWAS phenotype-specific SNPs and hypothesized that they would contribute to individual-level differences in both clinical and neuropathological traits within the ALS-FTD clinicopathologic spectrum, where individuals have heterogeneous clinical, neuropathological, and genetic traits. Cluster-specific PRS were constructed for C1 ALS and C2 FTLD-TDP as the weighted sum of alleles per SNP per cluster for each individual, centered and mean scaled. The difference between an individual’s C2 FTLD-TDP risk and C1 ALS risk was evaluated to assess relative FTLD-TDP to ALS risk.

Polygenic risk associates with clinical traits within mutation carriers at risk for neuromuscular and cognitive-behavioral impairment.

To assess whether relative polygenic risk was associated with clinical trait heterogeneity across the ALS-FTD spectrum, the difference between C2 FTLD-TDP risk and C1 ALS risk was tested for association with clinical traits. A one-way ANOVA revealed that C2 FTLD-TDP risk, relative to C1 ALS risk, was higher in individuals with cognitive-behavioral impairment than those with MND, controlling for last contact age, sex, and inclusion in the prior ALS or FTLD-TDP GWAS (F(2,248)=[7.18], p < 0.001), S5 Fig). Pairwise t-tests revealed an FDR-adjusted significant difference in the relative polygenic risk between the Cog-only and MND-only groups. This suggests that genetic variation associates with individual-level differences in clinical traits.

As our sample contained both sporadic and familial cases, polygenic risk was subsequently specifically tested for association with clinical trait heterogeneity within known C9orf72, GRN, TARDBP, TBK1, and VCP mutation carriers, where autosomal dominant inherited mutations may primarily drive disease more than underlying polygenic risk. As the Cog + MND group did not differ in relative risk from either the Cog-only or the MND-only group in the full sample, we tested cluster-specific polygenic risk against its corresponding trait, cognitive-behavioral impairment or neuromuscular impairment. One-way ANOVAs revealed that the C1 ALS PRS was higher in mutation carriers with MND (MND-only or Cog + MND, compared to Cog-only; F(1,125)=[6.87], p = 0.01), while the C2 FTLD-TDP PRS was higher in mutation carriers with cognitive-behavioral impairment (Cog-only or Cog + MND, compared to MND-only; (F(1,125)=[12.47], p < 0.001), controlling for last contact age, sex, and inclusion in the prior ALS or FTLD-TDP GWAS, respectively (S6 Fig). This suggests that even in the context of autosomal dominant disease, genetic variation associates with risk of a specific impairment including C1 ALS risk for neuromuscular impairment and C2 FTLD-TDP risk for cognitive-behavioral impairment.

Further limiting the sample to C9orf72 cases which can develop ALS, FTLD-TDP, or both, a one-way ANOVA revealed that the C1 ALS PRS was higher in C9orf72 carriers with MND (MND-only or Cog + MND, compared to Cog-only (F(1,104)=[4.2], p = 0.04), while the C2 FTLD-TDP PRS was higher in C9orf72 carriers with cognitive-behavioral impairment (Cog-only or Cog + MND, compared to MND-only; F(1,104)=[8.62], p = 0.004), Fig 3), controlling for last contact age, sex, and inclusion in the prior ALS or FTLD-TDP GWAS, respectively. This provides individual-level evidence that each cluster independently contributes to risk for a neuromuscular impairment or a cognitive-behavioral impairment in the context of a C9orf72 expansion.

thumbnail
Download:
Fig 3. C1 ALS and C2 FTLD-TDP associate with individual-level clinical traits in C9orf72 cases.

The (A) C1 ALS and (B) C2 FTLD-TDP PRS were tested for association with neuromuscular or cognitive-behavioral impairment or in C9orf72 mutation carriers.

https://doi.org/10.1371/journal.pone.0338398.g003

Polygenic risk associates with regional neuropathological traits.

TDP-43 inclusions are observed in characteristic anatomic regions correlating with clinical traits. To assess whether relative polygenic risk was associated with regional neuropathological traits across the ALS-FTD spectrum, the difference between C2 FTLD-TDP risk and C1 ALS risk was tested for association with the burden of TDP-43 pathology across brain regions. First, this was done in characteristic FTLD-TDP regions, specifically the cingulate gyrus, angular gyrus, middle frontal cortex, and superior/middle temporal cortex (S7 Fig). One-way ANOVAs revealed that higher C2 FTLD-TDP risk, relative to C1 ALS risk, was associated with higher TDP-43 burden in each FTLD-TDP characteristic region (all FDR adjusted p < .05, Fig 4), controlling for age at death, sex, and inclusion in the prior ALS or FTLD-TDP GWAS. Second, we tested whether relative polygenic risk was associated with TDP-43 burden in characteristic ALS regions, motor cortex and spinal cord. However, no significant differences in relative PRS by TDP-43 burden were observed in either ALS characteristic region. These observations suggest that these data-driven cluster-specific polygenic risk scores relate to characteristic individual-level differences in the anatomic distribution of TDP-43 within the ALS-FTD clinicopathologic spectrum.

thumbnail
Download:
Fig 4. Relative polygenic risk associates with regional neuropathological traits.

The difference between an individual’s C2 FTLD-TDP risk and C1 ALS risk was tested for association with the burden of TDP-43 pathology in FTLD-TDP characteristic regions (top row: cingulate gyrus, angular gyrus, middle frontal cortex, and superior/middle temporal cortex), and ALS characteristic regions (bottom row: motor cortex and spinal cord). The FDR adjusted p-values for each ANOVA are displayed by region. Asterisks represent FDR-adjusted significant pairwise t-tests within region.

https://doi.org/10.1371/journal.pone.0338398.g004

Discussion

In this study, we hypothesized that both shared and disparate genetic variants would be identified across TDP-43 proteinopathies which contribute to the heterogenous individual-level clinical and neuropathological traits of ALS and FTLD-TDP. We identified data-driven clusters of correlated SNPs that were either GWAS phenotype specific (C1 ALS, C2 FTLD-TDP, and C3 HS-Aging) or shared across TDP-43 proteinopathy GWAS phenotypes (C4 ALS-FTLD and C5 FTLD-HS). Leveraging commonly genotyped SNPs across existing independent GWAS allowed for the identification of patterns of correlations between SNPs and the respective GWAS phenotypes. Though performed at the summary statistic level, this analysis suggested GWAS phenotype-specific cluster SNPs which were then shown to associate with clinical and neuropathological traits at the individual level, including neuromuscular impairment and/or cognitive-behavioral impairment and the burden of TDP-43 pathology across brain regions.

Gene ontology enrichment analysis of cluster SNPs provided plausible reinforcement for the presence of both shared and GWAS phenotype-specific biological pathways across TDP-43 proteinopathies. Terms relevant to nervous system development, such as neuron projection development and morphogenesis, as well as those related to synapse structure, organization, activity, and assembly were common across C1 ALS, C2 FTLD-TDP, and C3 HS-Aging, as would be expected in neurodegenerative diseases. However, top overrepresented pathways also support disease-specific drivers. Several of the top overrepresented pathways in C1 ALS related to the selective vulnerability of motor neurons and spinal cord in ALS. Among these were terms related to motor neuron migration, including spinal cord motor neuron migration and lateral motor column neuron migration. This analysis also identified several terms in C1 ALS related to axonogenesis, including central nervous system projection neuron axonogenesis and specifically spinal cord ventral commissure morphogenesis, as well as anterior/posterior axon guidance and negative regulation of collateral sprouting. Further, detailed molecular studies have identified inflammatory responses following neurodegeneration in ALS [46,47], and have implicated inflammatory pathways as contributors to further neurodegeneration [48,49]. Gene expression studies in ALS have identified an upregulation of immune and inflammatory genes in disease relevant regions [50–53]. Here, we observed pathways related to the response to cytokines in C1 ALS, including gene ontology terms such as response to stem cell factor, cellular response to stem cell factor stimulus, cytokine-mediated signaling pathway, Kit signaling pathway, response to chemokine, and the regulation of chemokine-mediated signaling pathway. The top overrepresented pathway in C1 ALS was positive regulation of retrograde transport, endosome to Golgi. Golgi changes are well documented in ALS and one of the most consistent pathologic findings in motor neuron disease [54–56]. Finally, C1 ALS included terms related to the regulation of metabolic process, including regulation of kinase activity, regulation of hexokinase activity, and regulation of glucokinase activity. Several lines of evidence support the role of kinases in ALS pathogenesis, including the identification of kinase-encoding genes such as TBK1, mutations in which can cause ALS [57]. In C2 FTLD-TDP, several of the top overrepresented pathways related to cytoskeleton organization, including actin filament network formation, actin filament bundle organization, and actin filament bundle assembly. Differential expression of genes relating to the structure and function of the cytoskeleton have been observed in FTLD-TDP, and the cytoskeleton has been implicated more broadly in aging and neurodegeneration [58–61]. For C3 HS-Aging, we observed several overrepresented terms related to neurogenesis and neuron development, including neuron recognition, which was among the top overrepresented pathways for this cluster. As the hippocampus is the site of adult neurogenesis, the inclusion of these pathways align with the selective vulnerability of the hippocampus in HS-aging [62,63].

Within the ALS-FTD spectrum, the genetic architecture is complex. In addition to shared causal mutations, such as the C9orf72 repeat expansion, shared significant associations have been identified in UNC13A [20,23,24]. Similarly, TMEM106B has been implicated for risk in both FTLD-TDP and HS-Aging [3,11,12]. Combined studies across TDP-43 proteinopathies have largely focused on identifying shared genetic risk factors that would explain susceptibility across GWAS phenotypes through a common underlying biological mechanism. In contrast, the identification of GWAS phenotype-specific cluster SNPs (e.g., C1 ALS and C2 FTLD-TDP), excluding those SNPs found to be shared across phenotypes (e.g., C4 ALS-FTLD), allowed us to demonstrate the contribution of genetic variants to individual-level clinical and neuropathological traits through the construction of cluster-specific PRS. Specifically, PRS analyses showed that higher C2 FTLD-TDP risk, relative to C1 ALS risk, was associated with higher TDP-43 burden in FTLD-TDP characteristic regions including cingulate gyrus, angular gyrus, middle frontal cortex, and superior/middle temporal cortex. Orbito-frontal regions are affected early in FTLD-TDP, even when the overall burden of pathology is lower, though more widespread cortical pathology is observed in advanced cases [64]. In contrast, these regions are either spared or affected later in the course of ALS, where pathology spreads from motor neurons [65] While no significant differences in relative PRS by TDP-43 burden were observed in either ALS characteristic region evaluated, it is possible that this is confounded by neuron loss, particularly for the spinal cord where neuronal dropout can lower TDP-43 burden. Our evidence for genetic overlap through C4 ALS-FTLD and C5 FTLD-HS is consistent with phenotypic spectrum observed in disease along the ALS-FTD spectrum and LATE and FTLD impacting similar limbic structures. However, we did not observe a cluster associated with ALS-HS which is consistent with less reported clinical overlap between ALS and LATE, but requires future investigation.

It is currently unknown why some individuals with TDP-43 pathology develop neuromuscular impairments while others develop cognitive-behavioral impairments. Both the C1 ALS PRS and the C2 FTLD-TDP PRS showed associations with their respective expected clinical traits, even within known mutation carriers where disease may be less influenced by underlying polygenic risk. Strikingly, even within individuals with a C9orf72 repeat expansion, for which there is no current way to predict the development of ALS, FTLD-TDP, or both, the C1 ALS PRS and C2 FTLD-TDP PRS associated with individual-level clinical traits. While further validation is required, this observation has important implications for predictive genetic testing and management of clinical care, including patient monitoring. Moreover, current clinical trials focus on either ALS or FTLD-TDP populations in the absence of knowledge about individual risk. These novel PRS may therefore be important for prognostication, by more precisely predicting whether a given individual with a C9orf72 repeat expansion is more likely to develop neuromuscular or cognitive-behavioral impairments, and stratification, by identifying differential outcomes within subgroups, in the context of interpreting outcomes in a combined trial [66–69].

This study is limited by the complexity of diagnosis across TDP-43 proteinopathies. A more complete clinical evaluation of neuromuscular impairment in suspected FTLD-TDP cases and cognitive-behavioral impairment in ALS cases is needed to better characterize these individuals. Furthermore, clinical heterogeneity is present even within TDP-43 proteinopathies. For example, initial FTD symptoms can include changes in behavior or language, and initial ALS symptoms can begin in different locations (e.g., bulbar or limb). The relationship between cluster-specific polygenic scores and the variability within these clinical traits should be explored.

Additional challenges to complete, accurate diagnoses in the FTD-ALS spectrum include a lack of TDP-43 biomarkers and often rapid progression which may preclude evaluation of the full clinical picture. Thus, there is a reliance on autopsy. However, the regional distribution of pathology alone cannot determine the presence of a secondary neuromuscular or cognitive-behavioral impairments within FTLD-TDP or ALS. Therefore, our individual-level analyses focused on clinical traits (neuromuscular impairment and/or cognitive-behavioral impairment) and neuropathological traits (burden of TDP-43 pathology across brain regions) rather than separate diagnostic groups.

Further, even though the vast majority of HS-Aging cases have TDP-43 pathology, the pathological definition does not require the presence TDP-43. The more recently defined Limbic-predominant age-related TDP-43 encephalopathy (LATE) relies only on the presence of TDP-43 in the amygdala and subsequently hippocampus and middle frontal gyrus [70]. As hippocampal sclerosis is neither sufficient nor necessary for a LATE diagnosis, future work including a LATE GWAS is needed to disentangle the genotype associations with HS-aging from limbic TDP-43. In the present study we were unable to evaluate the relationship between C3 HS-Aging SNPs and individual-level differences in either clinical or neuropathological traits due to a lack of directionality information in the published GWAS summary statistics.

The threshold for inclusion of SNPs initially selected for analysis from the GWAS summary statistics was arbitrary. This threshold (the top 1% of SNPs with the strongest associations to any GWAS phenotype) is lenient by GWAS standards, including SNPs with summary statistic p-values as high as 0.003. However, strict GWAS correction would leave too few SNPs to compare across GWAS phenotypes, and sub-threshold associations are still likely informative. However, relaxing the threshold beyond a certain point would introduce non-meaningful SNPs which have no associations with any of the GWAS phenotypes. Importantly, we show that the SNPs chosen at this threshold are biologically meaningful with converging evidence from our summary statistic and individual level results.

Finally, similar to limitations of the vast majority of PRS studies, the cluster-specific polygenic scores were both predominantly derived from and tested on individuals of European ancestry. Therefore, future work is required to validate the broader applicability of these findings in a diverse population.

Overall, examining summary statistics revealed clusters of both shared and disparate correlated SNPs across GWAS phenotypes associated with TDP-43 proteinopathies. These clusters have biologically plausible pathway associations, and cluster SNPs relate to individual-level clinical and neuropathological traits providing suggestive evidence of a potentially predictive and prognostic genetic marker of neuromuscular and/or cognitive-behavioral impairment risk. We suggest that there are polygenic modifiers of clinical and neuropathological traits across TDP-43 proteinopathies that may contribute to individual-level differences, including the likelihood for developing FTLD or ALS.

Supporting information

S1 Fig. The top 1% of SNPs with the strongest associations to any GWAS phenotype were selected for analysis.

494,417 unique SNPs commonly genotyped across the ALS, FTLD-TDP, and HS-Aging GWAS summary statistics were ranked according to each SNP’s strongest association to any GWAS phenotype (by |z|). For each GWAS, (A) ALS, (B) FTLD-TDP, (C) HS-Aging, all SNPs are plotted by -log10(p-value) from the original GWAS and the percentile rank. The dashed line indicates the cut point for selection.

https://doi.org/10.1371/journal.pone.0338398.s001

(TIF)

S2 Fig. Cluster assignment is robust for SNPs associated with a single GWAS phenotype.

To test for cluster robustness, we created 100 samples, each containing a random 80% of the 4,945 SNPs selected for weighted correlation analysis, without replacement. For each resample, WGCNA was performed as in the full sample. Each of the 4,945 SNPs is represented on the x-axis, faceted by cluster assignment in the full sample. Each row on the y-axis is a resample, and the colors indicate which clusters the SNPs have been assigned in that resample. All resamples have 20% NAs by definition (white). Light and dark gray represent an assigned cluster 6 or 7, assigned in 31 and 1 resamples respectively, which does not exist in the full sample. In the majority of instances (n = 16 resamples), cluster 6 contained SNPs from only a single cluster, indicating the split of a single cluster into 2 clusters. As is visually apparent, clusters associated with a single GWAS phenotype, C1 ALS, C2 FTLD-TDP, and C3 HS-Aging, are particularly robust, with 99.76%, 99.63%, and 97.9% SNPs, respectively, assigned to the same cluster as in the full sample across 100 resamples. In each resample, cluster numbers are assigned by size of the cluster. Importantly, while C4 ALS-FTLD and C5 FTLD-HS SNPs are assigned to both clusters 4 and 5 across the resamples, there is only one instance of a resample in which C5 FTLD-HS SNPs are assigned to both cluster 4 and 5. Thus, the separation of C4 ALS-FTLD from C5 FTLD-HS SNPs is maintained. 73.98% of C4 ALS-FTLD and 85.31% of C5 FTLD-HS SNPs were assigned to either the cluster 4 or 5 across all resamples. The largest differences in C4 ALS-FTLD and C5 FTLD-HS SNP assignments in the resamples compared to the full sample assignment came from 21.41% of C4 ALS-FTLD SNPs assigned to cluster 1 and 7.77% of C5 FTLD-HS SNPs assigned to cluster 2. The reduced robustness in the clusters representing phenotypic pairs (C4 ALS-FTLD and C5 FTLD-HS) compared to a single GWAS phenotype (C1 ALS, C2 FTLD-TDP, and C3 HS-Aging), is likely influenced by their size. Single GWAS phenotype clusters are an order of magnitude larger than phenotypic pair clusters (1042–1910 vs 121–144 SNPs per cluster, respectively).

https://doi.org/10.1371/journal.pone.0338398.s002

(TIF)

S3 Fig. Cluster SNPs were pruned by LD.

Cluster SNPs are plotted by relative associations (by |z|) to (A) FTLD-TDP and ALS, (B) HS-Aging and ALS, or (C) HS-Aging and FTLD-TDP GWAS. SNPs were LD pruned within clusters using an R2 threshold of 0.3. SNPs shown in black were carried forward for analysis, while those in gray were removed.

https://doi.org/10.1371/journal.pone.0338398.s003

(TIF)

S4 Fig. Common pathways between clusters are relevant to nervous system processes.

Of 382 unique pathways significant across clusters, 76 were common to C1 ALS and C2 FTLD-TDP, 57 were common to C1 ALS and C3 HS-Aging, and 52 were common to C2 FTLD-TDP and C3 HS-Aging. All common pathways are plotted by fold enrichment within (A) C1 ALS and C2 FTLD-TDP, (B) C1 ALS and C3 HS-Aging, or (C) C2 FTLD-TDP and C3 HS-Aging. Terms in red contain neruo- (neuron, neurogenesis, neurotransmitter), nervous system, synap- (synapse, synaptic), dendrit- (dendrite, dendritic), or axon.

https://doi.org/10.1371/journal.pone.0338398.s004

(TIF)

S5 Fig. Relative polygenic risk associates with individual-level clinical traits.

The difference between an individual’s C2 FTLD-TDP risk and C1 ALS risk was tested for association with neuromuscular or cognitive-behavioral impairment. Asterisk represents FDR-adjusted significant pairwise t-test.

https://doi.org/10.1371/journal.pone.0338398.s005

(TIF)

S6 Fig. C1 ALS and C2 FTLD-TDP associate with individual-level clinical traits in known mutation carriers.

The (A) C1 ALS and (B) C2 FTLD-TDP PRS were tested for association with neuromuscular or cognitive-behavioral impairment in C9orf72, GRN, TARDBP, TBK1, and VCP mutation carriers.

https://doi.org/10.1371/journal.pone.0338398.s006

(TIF)

S7 Fig. Regional burden of TDP-43 in characteristic FTLD-TDP and ALS regions.

Ordinal ratings of TDP-43 burden are displayed for each individual, grouped based on clinical evidence for neuromuscular impairment (MND-only), cognitive and/or behavioral impairment (Cog-only), or any evidence of a combination of impairment (Cog + MND) at any point in the natural history of disease.

https://doi.org/10.1371/journal.pone.0338398.s007

(TIF)

S1 File. Data-driven clusters of correlated SNPs.

SNPs that remain after pruning are listed for each cluster.

https://doi.org/10.1371/journal.pone.0338398.s008

(PDF)

S2 File. Gene ontology enrichment analysis results by cluster.

Pathways that met FDR-adjusted p < .05 are listed for each cluster.

https://doi.org/10.1371/journal.pone.0338398.s009

(PDF)

S3 File. C1 ALS PRS SNPs.

https://doi.org/10.1371/journal.pone.0338398.s010

(PDF)

S4 File. C2 FTLD-TDP PRS SNPs.

https://doi.org/10.1371/journal.pone.0338398.s011

(PDF)

References

  1. 1. Neumann M, Sampathu DM, Kwong LK, Truax AC, Micsenyi MC, Chou TT, et al. Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Science. 2006;314(5796):130–3. pmid:17023659
  2. 2. Amador-Ortiz C, Ahmed Z, Zehr C, Dickson DW. Hippocampal sclerosis dementia differs from hippocampal sclerosis in frontal lobe degeneration. Acta Neuropathol. 2007;113(3):245–52. pmid:17195931
  3. 3. Nelson PT, Estus S, Abner EL, Parikh I, Malik M, Neltner JH, et al. ABCC9 gene polymorphism is associated with hippocampal sclerosis of aging pathology. Acta Neuropathol. 2014;127(6):825–43. pmid:24770881
  4. 4. Nag S, Yu L, Capuano AW, Wilson RS, Leurgans SE, Bennett DA, et al. Hippocampal sclerosis and TDP-43 pathology in aging and Alzheimer disease. Ann Neurol. 2015;77(6):942–52. pmid:25707479
  5. 5. Strong MJ, Abrahams S, Goldstein LH, Woolley S, Mclaughlin P, Snowden J, et al. Amyotrophic lateral sclerosis - frontotemporal spectrum disorder (ALS-FTSD): Revised diagnostic criteria. Amyotroph Lateral Scler Frontotemporal Degener. 2017;18(3–4):153–74. pmid:28054827
  6. 6. Dickson DW, Davies P, Bevona C, Van Hoeven KH, Factor SM, Grober E, et al. Hippocampal sclerosis: a common pathological feature of dementia in very old (> or = 80 years of age) humans. Acta Neuropathol. 1994;88(3):212–21. pmid:7810292
  7. 7. Corey-Bloom J, Sabbagh MN, Bondi MW, Hansen L, Alford MF, Masliah E, et al. Hippocampal sclerosis contributes to dementia in the elderly. Neurology. 1997;48(1):154–60. pmid:9008511
  8. 8. Leverenz JB, Agustin CM, Tsuang D, Peskind ER, Edland SD, Nochlin D, et al. Clinical and neuropathological characteristics of hippocampal sclerosis: a community-based study. Arch Neurol. 2002;59(7):1099–106. pmid:12117357
  9. 9. Nicolas A, Kenna KP, Renton AE, Ticozzi N, Faghri F, Chia R, et al. Genome-wide Analyses Identify KIF5A as a Novel ALS Gene. Neuron. 2018;97(6):1267–88. pmid:29566793
  10. 10. van Es MA, Veldink JH, Saris CGJ, Blauw HM, van Vught PWJ, Birve A, et al. Genome-wide association study identifies 19p13.3 (UNC13A) and 9p21.2 as susceptibility loci for sporadic amyotrophic lateral sclerosis. Nat Genet. 2009;41(10):1083–7. pmid:19734901
  11. 11. Van Deerlin VM, Sleiman PMA, Martinez-Lage M, Chen-Plotkin A, Wang L-S, Graff-Radford NR, et al. Common variants at 7p21 are associated with frontotemporal lobar degeneration with TDP-43 inclusions. Nat Genet. 2010;42(3):234–9. pmid:20154673
  12. 12. Beecham GW, Hamilton K, Naj AC, Martin ER, Huentelman M, Myers AJ, et al. Genome-wide association meta-analysis of neuropathologic features of Alzheimer’s disease and related dementias. PLoS Genet. 2014;10(9):e1004606. pmid:25188341
  13. 13. van Rheenen W, Shatunov A, Dekker AM, McLaughlin RL, Diekstra FP, Pulit SL, et al. Genome-wide association analyses identify new risk variants and the genetic architecture of amyotrophic lateral sclerosis. Nat Genet. 2016;48(9):1043–8. pmid:27455348
  14. 14. Finch N, Carrasquillo MM, Baker M, Rutherford NJ, Coppola G, Dejesus-Hernandez M, et al. TMEM106B regulates progranulin levels and the penetrance of FTLD in GRN mutation carriers. Neurology. 2011;76(5):467–74. pmid:21178100
  15. 15. Gallagher MD, Suh E, Grossman M, Elman L, McCluskey L, Van Swieten JC, et al. TMEM106B is a genetic modifier of frontotemporal lobar degeneration with C9orf72 hexanucleotide repeat expansions. Acta Neuropathol. 2014;127(3):407–18. pmid:24442578
  16. 16. Nicholson AM, Finch NA, Wojtas A, Baker MC, Perkerson RB 3rd, Castanedes-Casey M, et al. TMEM106B p.T185S regulates TMEM106B protein levels: implications for frontotemporal dementia. J Neurochem. 2013;126(6):781–91. pmid:23742080
  17. 17. Zhou X, Nicholson AM, Ren Y, Brooks M, Jiang P, Zuberi A, et al. Loss of TMEM106B leads to myelination deficits: implications for frontotemporal dementia treatment strategies. Brain. 2020;143(6):1905–19. pmid:32504082
  18. 18. van Blitterswijk M, Mullen B, Nicholson AM, Bieniek KF, Heckman MG, Baker MC, et al. TMEM106B protects C9ORF72 expansion carriers against frontotemporal dementia. Acta Neuropathol. 2014;127(3):397–406. pmid:24385136
  19. 19. van der Zee J, Van Langenhove T, Kleinberger G, Sleegers K, Engelborghs S, Vandenberghe R, et al. TMEM106B is associated with frontotemporal lobar degeneration in a clinically diagnosed patient cohort. Brain. 2011;134(Pt 3):808–15. pmid:21354975
  20. 20. Diekstra FP, Van Deerlin VM, van Swieten JC, Al-Chalabi A, Ludolph AC, Weishaupt JH, et al. C9orf72 and UNC13A are shared risk loci for amyotrophic lateral sclerosis and frontotemporal dementia: a genome-wide meta-analysis. Ann Neurol. 2014;76(1):120–33. pmid:24931836
  21. 21. Pottier C, Ren Y, Perkerson RB 3rd, Baker M, Jenkins GD, van Blitterswijk M, et al. Genome-wide analyses as part of the international FTLD-TDP whole-genome sequencing consortium reveals novel disease risk factors and increases support for immune dysfunction in FTLD. Acta Neuropathol. 2019;137(6):879–99. pmid:30739198
  22. 22. Karch CM, Wen N, Fan CC, Yokoyama JS, Kouri N, Ross OA, et al. Selective Genetic Overlap Between Amyotrophic Lateral Sclerosis and Diseases of the Frontotemporal Dementia Spectrum. JAMA Neurol. 2018;75(7):860–75. pmid:29630712
  23. 23. Renton AE, Majounie E, Waite A, Simón-Sánchez J, Rollinson S, Gibbs JR, et al. A hexanucleotide repeat expansion in C9ORF72 is the cause of chromosome 9p21-linked ALS-FTD. Neuron. 2011;72(2):257–68. pmid:21944779
  24. 24. DeJesus-Hernandez M, Mackenzie IR, Boeve BF, Boxer AL, Baker M, Rutherford NJ, et al. Expanded GGGGCC hexanucleotide repeat in noncoding region of C9ORF72 causes chromosome 9p-linked FTD and ALS. Neuron. 2011;72(2):245–56. pmid:21944778
  25. 25. Brooks BR. El Escorial World Federation of Neurology criteria for the diagnosis of amyotrophic lateral sclerosis. Subcommittee on Motor Neuron Diseases/Amyotrophic Lateral Sclerosis of the World Federation of Neurology Research Group on Neuromuscular Diseases and the El Escorial “Clinical limits of amyotrophic lateral sclerosis” workshop contributors. J Neurol Sci. 1994;124 Suppl:96–107. pmid:7807156
  26. 26. Hinrichs AS, Karolchik D, Baertsch R, Barber GP, Bejerano G, Clawson H, et al. The UCSC Genome Browser Database: update 2006. Nucleic Acids Res. 2006;34(Database issue):D590-8. pmid:16381938
  27. 27. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008;9:559. pmid:19114008
  28. 28. Myers TA, Chanock SJ, Machiela MJ. LDlinkR: An R Package for Rapidly Calculating Linkage Disequilibrium Statistics in Diverse Populations. Front Genet. 2020;11:157. pmid:32180801
  29. 29. Durinck S, Moreau Y, Kasprzyk A, Davis S, De Moor B, Brazma A, et al. BioMart and Bioconductor: a powerful link between biological databases and microarray data analysis. Bioinformatics. 2005;21(16):3439–40. pmid:16082012
  30. 30. Kolberg L, Raudvere U, Kuzmin I, Vilo J, Peterson H. gprofiler2 -- an R package for gene list functional enrichment analysis and namespace conversion toolset g:Profiler. F1000Res. 2020;9:ELIXIR–709. pmid:33564394
  31. 31. Xie SX, Baek Y, Grossman M, Arnold SE, Karlawish J, Siderowf A, et al. Building an integrated neurodegenerative disease database at an academic health center. Alzheimers Dement. 2011;7(4):e84–93. pmid:21784346
  32. 32. Das S, Forer L, Schönherr S, Sidore C, Locke AE, Kwong A, et al. Next-generation genotype imputation service and methods. Nat Genet. 2016;48(10):1284–7. pmid:27571263
  33. 33. Taliun D, Harris DN, Kessler MD, Carlson J, Szpiech ZA, Torres R, et al. Sequencing of 53,831 diverse genomes from the NHLBI TOPMed Program. Nature. 2021;590(7845):290–9. pmid:33568819
  34. 34. Zhao S, Jing W, Samuels DC, Sheng Q, Shyr Y, Guo Y. Strategies for processing and quality control of Illumina genotyping arrays. Brief Bioinform. 2018;19(5):765–75. pmid:28334151
  35. 35. Suh E, Lee EB, Neal D, Wood EM, Toledo JB, Rennert L, et al. Semi-automated quantification of C9orf72 expansion size reveals inverse correlation between hexanucleotide repeat number and disease duration in frontotemporal degeneration. Acta Neuropathol. 2015;130(3):363–72. pmid:26022924
  36. 36. Toledo JB, Van Deerlin VM, Lee EB, Suh E, Baek Y, Robinson JL, et al. A platform for discovery: The University of Pennsylvania Integrated Neurodegenerative Disease Biobank. Alzheimers Dement. 2014;10(4):477-484.e1. pmid:23978324
  37. 37. Rosen DR. Mutations in Cu/Zn superoxide dismutase gene are associated with familial amyotrophic lateral sclerosis. Nature. 1993;364(6435):362. pmid:8332197
  38. 38. Freischmidt A, Wieland T, Richter B, Ruf W, Schaeffer V, Müller K, et al. Haploinsufficiency of TBK1 causes familial ALS and fronto-temporal dementia. Nat Neurosci. 2015;18(5):631–6. pmid:25803835
  39. 39. Kabashi E, Valdmanis PN, Dion P, Spiegelman D, McConkey BJ, Vande Velde C, et al. TARDBP mutations in individuals with sporadic and familial amyotrophic lateral sclerosis. Nat Genet. 2008;40(5):572–4. pmid:18372902
  40. 40. Johnson JO, Mandrioli J, Benatar M, Abramzon Y, Van Deerlin VM, Trojanowski JQ, et al. Exome sequencing reveals VCP mutations as a cause of familial ALS. Neuron. 2010;68(5):857–64. pmid:21145000
  41. 41. Cruts M, Gijselinck I, van der Zee J, Engelborghs S, Wils H, Pirici D, et al. Null mutations in progranulin cause ubiquitin-positive frontotemporal dementia linked to chromosome 17q21. Nature. 2006;442(7105):920–4. pmid:16862115
  42. 42. Baker M, Mackenzie IR, Pickering-Brown SM, Gass J, Rademakers R, Lindholm C, et al. Mutations in progranulin cause tau-negative frontotemporal dementia linked to chromosome 17. Nature. 2006;442(7105):916–9. pmid:16862116
  43. 43. Mackenzie IRA, Neumann M, Baborie A, Sampathu DM, Du Plessis D, Jaros E, et al. A harmonized classification system for FTLD-TDP pathology. Acta Neuropathol. 2011;122(1):111–3. pmid:21644037
  44. 44. Mackenzie IRA, Neumann M, Bigio EH, Cairns NJ, Alafuzoff I, Kril J, et al. Nomenclature and nosology for neuropathologic subtypes of frontotemporal lobar degeneration: an update. Acta Neuropathol. 2010;119(1):1–4. pmid:19924424
  45. 45. Lee EB, Porta S, Michael Baer G, Xu Y, Suh E, Kwong LK, et al. Expansion of the classification of FTLD-TDP: distinct pathology associated with rapidly progressive frontotemporal degeneration. Acta Neuropathol. 2017;134(1):65–78. pmid:28130640
  46. 46. Town T, Nikolic V, Tan J. The microglial “activation” continuum: from innate to adaptive responses. J Neuroinflammation. 2005;2:24. pmid:16259628
  47. 47. Liddelow SA, Guttenplan KA, Clarke LE, Bennett FC, Bohlen CJ, Schirmer L, et al. Neurotoxic reactive astrocytes are induced by activated microglia. Nature. 2017;541(7638):481–7. pmid:28099414
  48. 48. Zhao W, Xie W, Le W, Beers DR, He Y, Henkel JS, et al. Activated microglia initiate motor neuron injury by a nitric oxide and glutamate-mediated mechanism. J Neuropathol Exp Neurol. 2004;63(9):964–77. pmid:15453095
  49. 49. Haidet-Phillips AM, Hester ME, Miranda CJ, Meyer K, Braun L, Frakes A, et al. Astrocytes from familial and sporadic ALS patients are toxic to motor neurons. Nat Biotechnol. 2011;29(9):824–8. pmid:21832997
  50. 50. D’Erchia AM, Gallo A, Manzari C, Raho S, Horner DS, Chiara M, et al. Massive transcriptome sequencing of human spinal cord tissues provides new insights into motor neuron degeneration in ALS. Sci Rep. 2017;7(1):10046. pmid:28855684
  51. 51. Brohawn DG, O’Brien LC, Bennett JP Jr. RNAseq Analyses Identify Tumor Necrosis Factor-Mediated Inflammation as a Major Abnormality in ALS Spinal Cord. PLoS One. 2016;11(8):e0160520. pmid:27487029
  52. 52. Dols-Icardo O, Montal V, Sirisi S, López-Pernas G, Cervera-Carles L, Querol-Vilaseca M, et al. Motor cortex transcriptome reveals microglial key events in amyotrophic lateral sclerosis. Neurol Neuroimmunol Neuroinflamm. 2020;7(5):e829. pmid:32669313
  53. 53. Andrés-Benito P, Moreno J, Aso E, Povedano M, Ferrer I. Amyotrophic lateral sclerosis, gene deregulation in the anterior horn of the spinal cord and frontal cortex area 8: implications in frontotemporal lobar degeneration. Aging (Albany NY). 2017;9(3):823–51. pmid:28283675
  54. 54. Soo KY, Halloran M, Sundaramoorthy V, Parakh S, Toth RP, Southam KA, et al. Rab1-dependent ER-Golgi transport dysfunction is a common pathogenic mechanism in SOD1, TDP-43 and FUS-associated ALS. Acta Neuropathol. 2015;130(5):679–97. pmid:26298469
  55. 55. Sundaramoorthy V, Sultana JM, Atkin JD. Golgi fragmentation in amyotrophic lateral sclerosis, an overview of possible triggers and consequences. Front Neurosci. 2015;9:400. pmid:26578862
  56. 56. Burk K, Pasterkamp RJ. Disrupted neuronal trafficking in amyotrophic lateral sclerosis. Acta Neuropathol. 2019;137(6):859–77. pmid:30721407
  57. 57. Guo W, Vandoorne T, Steyaert J, Staats KA, Van Den Bosch L. The multifaceted role of kinases in amyotrophic lateral sclerosis: genetic, pathological and therapeutic implications. Brain. 2020;143(6):1651–73. pmid:32206784
  58. 58. Chen-Plotkin AS, Geser F, Plotkin JB, Clark CM, Kwong LK, Yuan W, et al. Variations in the progranulin gene affect global gene expression in frontotemporal lobar degeneration. Hum Mol Genet. 2008;17(10):1349–62. pmid:18223198
  59. 59. Andrés-Benito P, Gelpi E, Povedano M, Santpere G, Ferrer I. Gene Expression Profile in Frontal Cortex in Sporadic Frontotemporal Lobar Degeneration-TDP. J Neuropathol Exp Neurol. 2018;77(7):608–27. pmid:29788265
  60. 60. Kounakis K, Tavernarakis N. The Cytoskeleton as a Modulator of Aging and Neurodegeneration. Adv Exp Med Biol. 2019;1178:227–45. pmid:31493230
  61. 61. Mol MO, Miedema SSM, van Swieten JC, van Rooij JGJ, Dopper EGP. Molecular Pathways Involved in Frontotemporal Lobar Degeneration with TDP-43 Proteinopathy: What Can We Learn from Proteomics? Int J Mol Sci. 2021;22(19):10298. pmid:34638637
  62. 62. Altman J. Are new neurons formed in the brains of adult mammals?. Science. 1962;135(3509):1127–8. pmid:13860748
  63. 63. Eriksson PS, Perfilieva E, Björk-Eriksson T, Alborn AM, Nordborg C, Peterson DA, et al. Neurogenesis in the adult human hippocampus. Nat Med. 1998;4(11):1313–7. pmid:9809557
  64. 64. Brettschneider J, Del Tredici K, Irwin DJ, Grossman M, Robinson JL, Toledo JB, et al. Sequential distribution of pTDP-43 pathology in behavioral variant frontotemporal dementia (bvFTD). Acta Neuropathol. 2014;127(3):423–39. pmid:24407427
  65. 65. Brettschneider J, Del Tredici K, Toledo JB, Robinson JL, Irwin DJ, Grossman M, et al. Stages of pTDP-43 pathology in amyotrophic lateral sclerosis. Ann Neurol. 2013;74(1):20–38. pmid:23686809
  66. 66. Spencer BE, Digma LA, Jennings RG, Brewer JB. Gene‐ and age‐informed enrichment method for preclinical Alzheimer’s disease trials: Approach and estimated cost savings. Alzheimer’s & Dementia. 2020;16(S5).
  67. 67. Broce IJ, Castruita PA, Yokoyama JS. Moving Toward Patient-Tailored Treatment in ALS and FTD: The Potential of Genomic Assessment as a Tool for Biological Discovery and Trial Recruitment. Front Neurosci. 2021;15:639078. pmid:33732107
  68. 68. Lu X, Liu Z, Cui Q, Liu F, Li J, Niu X, et al. A polygenic risk score improves risk stratification of coronary artery disease: a large-scale prospective Chinese cohort study. Eur Heart J. 2022;43(18):1702–11. pmid:35195259
  69. 69. Huntley C, Torr B, Sud A, Rowlands CF, Way R, Snape K, et al. Utility of polygenic risk scores in UK cancer screening: a modelling analysis. Lancet Oncol. 2023;24(6):658–68. pmid:37178708
  70. 70. Nelson PT, Dickson DW, Trojanowski JQ, Jack CR, Boyle PA, Arfanakis K, et al. Limbic-predominant age-related TDP-43 encephalopathy (LATE): consensus working group report. Brain. 2019;142(6):1503–27. pmid:31039256